Remote sensing of photosynthetic-light-use efficiency of boreal forest
Caroline J. Nichol
a,∗, Karl F. Huemmrich
b,1, T. Andrew Black
c,2, Paul G. Jarvis
a,3,
Charles L. Walthall
d,4, John Grace
a,5, Forrest G. Hall
b,6aInstitute of Ecology and Resource Management, University of Edinburgh, EH9 3JU, Scotland, UK bNASA Goddard Space Flight Center, Code 923, Greenbelt, MD 20771, USA
cDepartment of Agricultural Sciences, University of British Columbia, Vancouver, Canada V6T 1Z4 dUnited States Agricultural Research Service, Remote Sensing Laboratory, Beltsville, MD 20707, USA
Received 12 April 1999; received in revised form 22 November 1999; accepted 23 November 1999
Abstract
Using a helicopter-mounted portable spectroradiometer and continuous eddy covariance data we were able to evaluate the photochemical reflectance index (PRI) as an indicator of canopy photosynthetic light-use efficiency (LUE) in four boreal forest species during the Boreal Ecosystem Atmosphere experiment (BOREAS). PRI was calculated from narrow waveband reflectance data and correlated with LUE calculated from eddy covariance data. Significant linear correlations were found between PRI and LUE when the four species were grouped together and when divided into functional type: coniferous and deciduous. Data from the helicopter-mounted spectroradiometer were then averaged to represent data generated by the Airborne Visible Infrared Imaging Spectrometer (AVIRIS). We calculated PRI from these data and relationships with canopy LUE were investigated. The relationship between PRI and LUE was weakened for deciduous species but strengthened for the coniferous species. The robust nature of this relationship suggests that relative photosynthetic rates may be derived from remotely-sensed reflectance measurements. ©2000 Elsevier Science B.V. All rights reserved.
Keywords: Xanthophyll cycle; AVIRIS; Photochemical reflectance index; Eddy covariance
∗
Corresponding author. Tel.:+44-131-650-5744; fax:+44-131-662-0478.
E-mail addresses: [email protected] (C.J. Nichol), [email protected] (K.F. Huemmrich),
[email protected] (T.A. Black), [email protected] (P.G. Jarvis), [email protected] (C.L. Walthall), [email protected] (J. Grace), [email protected] (F.G. Hall).
1Tel.:+301-286-4862; fax:+301-286-0239. 2Tel.:+604-822-2730; fax:+604-822-8639. 3Tel.:+44-131-650-5426; fax:+44-131-662-0478. 4Tel.:+301-504-6074; fax:+301-504-5031. 5Tel.:+44-131-650-5400; fax:+44-131-662-0478. 6Tel.:+301-286-2974; Fax:+301-286-0239.
1. Introduction
Remote sensing offers an important landscape and regional perspective on vegetation structure and func-tion (Asner et al., 1998). Much remote sensing of veg-etation has made use of broad band sensors to derive indices of vegetation cover, such as the normalised difference vegetation index (NDVI) (Tucker, 1979). Although NDVI often correlates well with biomass, leaf area index (LAI) and fraction of absorbed photo-synthetically active radiation (FAPAR), it often fails to capture physiological processes that occur on fine temporal and spectral scales. For example, drought tolerant evergreens can undergo significant changes in photosynthetic light-use efficiency (LUE), without
an equivalent change in NDVI (Running and Nemani, 1988). Photosynthesis can change from day to day or hour to hour depending on the response of the leaves to a naturally fluctuating environment, with no signif-icant changes in canopy architecture or NDVI (Ten-hunen et al., 1987).
Pioneering work has shown how the reflectance of leaves may contain a signal for photosynthetic efficiency, and therefore provide a new method to detect changes in photosynthesis using remote
sens-ing. Photosynthetic LUE is defined here as CO2
uptake divided by the incident photosynthetic photon flux density (PPFD) and since the 1930s it has been known to decline at high PPFD as photosynthesis becomes light-saturated. More recently it has been found that when the photosynthetic system receives excess excitation energy from sunlight, the xantho-phyll cycle is affected. The carotenoid violaxanthin is converted to zeaxanthin via de-epoxidase reactions (Yamamoto, 1979). Recent experimental work has shown a close correlation between xanthophyll pig-ment interconversion and the dissipation of excess en-ergy in the pigment bed associated with photosystem II (PS II) (for reviews see Pfundel and Bilger, 1994; Demmig-Adams and Adams, 1994; Demmig-Adams et al., 1996). Recent evidence indicates that the con-version of the pigment violaxanthin to the photopro-tective pigment zeaxanthin acts to lower the energy level of the lowest excited singlet state below that of chlorophyll-a, thus providing a sink for the excess excitation energy (Frank et al., 1994; Owens, 1996).
The interconversion of the xanthophyll cycle pig-ments can be detected in leaves through a change in the absorbance at 505–515 nm (Bilger et al., 1989) or the reflectance at 531 nm (Gamon et al., 1990, 1993). Because the pigments of the xanthophyll cycle are so closely linked to the LUE of PS II, a reflectance index that incorporates reflectance at 531 nm could provide a remote indicator of photosynthetic function. Gamon et al. (1990) formulated the photochemical reflectance index (PRI), incorporating reflectance at 531 nm (the xanthophyll cycle signal) and a reference wavelength (Gamon et al., 1992).
A number of studies beginning in the late 1980s explored the relationships between PRI and photo-synthetic LUE at the leaf and small plot level. Mea-surements on individual leaves have demonstrated that PRI, calculated from narrow waveband data,
was closely related to1F/Fm′, a fluorescence-based
indicator of PSII LUE, as well as LUE calculated from gas exchange measurements in leaves from a wide range of species (Penuelas et al., 1995). Fillela et al. (1996) and Gamon et al. (1997) have further presented evidence that PRI provided a widely ap-plicable index of leaf LUE across species, functional types and nutrient levels.
The only canopy scale measurements that exist are on small plots (∼2 m2) and single species. Gamon et al. (1992) measured reflectance of sunflower plants with a portable spectroradiometer from a height of 4 m with wide-angle optics, allowing reflectance measure-ments over an area of diameter 1 m (consisting of 8–10 plants). This work showed that PRI closely tracked di-urnal changes in photosynthetic LUE in control and nitrogen stressed canopies, but not in water stressed canopies that were undergoing severe wilting (Gamon et al., 1992).
Such observations suggest that measurement of re-flectance by remote sensing in these spectral regions should also be useful for inferring canopy-scale LUE over regional and larger areas. However, at this time no work has been done to relate canopy PRI to LUE over areas of tens of meters. We have investigated the use of PRI as an index of LUE over contrasting canopy types. Extensive canopy reflectance and eddy covari-ance measurements of gas exchange were made over boreal forest during the BOREAS experiment (Boreal Ecosystem Atmosphere Study, for overview see Sell-ers et al., 1995) and provided the opportunity to inves-tigate the PRI:LUE relationship over larger (∼70 m2) forested areas.
We report in this paper an evaluation of PRI as an indicator or photosynthetic LUE. We begin by describ-ing the sites sampled in the boreal forest and the data set used, then proceed to outline the analysis approach for PRI:LUE. Finally, we present the results of this analysis and draw a number of conclusions regarding future directions.
2. Materials and methods
2.1. Study sites
Cana-dian boreal forest. A Northern Study Area (NSA) was located near Thomson, Manitoba, and Southern Study Area (SSA) was located near Candle Lake, Sask. (see Sellers et al., 1995b for details) and specific sub-sites within the SSA were intensively studied during the growing season. A brief outline of the sites used in this study is included here.
2.2. Old black spruce
The site lies in the southern mixed forest zone about 100 km north of Prince Albert in Sask. (53◦55′N,
105◦5′W, elevation 630 m), and has been described
by Jarvis et al. (1997). The terrain is essentially flat with pure and mixed stands of black (Picea mariana (Mill.) BSP) and white spruce (P. glauca Moench.), jack pine (Pinus banksiana Lamb.), aspen, fen and lakes. The substratum is peat overlying glacial drift with an elevated water table so the surface is gener-ally wet. The understorey is sparse with some low shrubs reaching∼1.5 m in height.
2.3. Old jack pine
The jack pine forest (Pinus banksiana Lamb.) stand grows in the SSA near Nipawin, Sask., Canada (53◦54′N, 104◦41′W, elevation 579 m) and has been
described by Baldocchi et al. (1997). The landscape
is relatively flat (slope ∼2 and 5%) with pure and
mixed stands of black (P. mariana (Mill.) BSP) and white spruce (P. glauca Moench.), aspen, jack pine and fen. The soil is a coarse textured, well drained sand. The ground was covered with an optically bright mat consisting of bearberry (Arctostaphylos uva-ursi), bog cranberry (Vaccinium vitis-idaea), and lichens (Cladina spp.).The understorey vegetation is sparse but there are isolated groups of alder (Alnus crispa).
2.4. Old aspen
The site is located in the SSA within Prince Albert National Park, Sask. (53◦38′N, 106◦12′W, elevation
600 m) and has been described by Hogg et al. (1997). It is an extensive, mostly pure, stand of trembling aspen (Populus tremuloides Michx.) Orthic gray luvisol with a loam texture dominates the site. Beaked hazelnut (Corylus cornutta Marsh) dominates the understorey
shrub layer and is approximately 2 m in height. Wild rose (Rosa woodsii) and alder (A. crispa) are found in intermittently.
2.5. Fen
The fen study site is located in the SSA, 115 km northeast of Prince Albert, Sask. (53◦57′N, 105◦57′W,
elevation 525 m). The fen study site is a minerotrophic, patterned fen surrounded by black spruce (P. mariana (Mill.) BSP)) and jack pine (Pinus banksiana Lamb.) forests and has been described by Sukyer et al. (1997). Abundant herbaceous species throughout the fen in-cluded bogbean (Menyanthes trifoliata) and several sedge species (Carex and Eriophorum spp.). Domi-nant woody plant species are 0.5–1.5 m tall bog birch (Betula pumila) and widely scattered, stunted tama-rack trees (Larix laricina). Mosses whilst present are not abundant in this fen.
2.6. Canopy spectral reflectance data
The National Aeronautics and Space Admin-istration (NASA) Goddard Space Flight Center (GSFC)/Wallops Flight Facility (WFF) helicopter-based optical remote sensing system (Walthall et al., 1996) was deployed to acquire canopy multispectral data with a portable spectroradiometer. A spectro-radiometer (model SE-590, Spectron Engineering, Denver, Colorado) was mounted on a steel rack at nadir orientation and processed to an atmospherically corrected at-surface reflectance. A 15◦ instantaneous
field of view lens (IFOV) was fitted to yield a ground resolution of 79 m at the 300 m nominal altitude. The SE-590 had a spectral range of 362.7–1122.7 nm, with a usable range of 400–900 nm. Data are reported at 3 nm intervals calculated at the centre point of a five 3 nm bin running average.
Table 1
Summary of the dates of the reflectance data acquisition with the range of temperature and solar radiation experienced during the observation periods
Site Instrument Observation date Range of temperature◦C Range of solar
radiationmmol m−2s−1
Old aspen SE-590 31-May-94 17.7–18.3 992–1129
21-July-94 22.6–23.3 1339–1539
25-July-94 21.5–21.7 1343–1450
13-Sept-94 15.7–16.4 1077–1216
Fen SE-590 01-June-94 20.5–21.0 1043–1352
06-June-94 21.2–21.2 1445–1545
21-July-94 21.3–21.6 1222–1366
13-Sept-94 12.4–12.9 1021–1117
Old jack pine SE-590 31-May-94 15.9–16.7 1022–1192
06-June-94 21.0–21.2 1416–1545
21-July-94 25.4–25.7 1397–1476
24-July-94 21.1–21.4 1382–1498
13-Sept-94 18.9–19.5 1196–1271
Old black spruce SE-590 01-June-94 18.4–18.6 1399–1498
06-June-94 20.4–20.6 1361–1479
21-July-94 24.6–25.3 1500–1579
23-July-94 18.5–18.9 1216–1341
13-Sept-94 18.7–19.1 1177–1236
to at-surface reflectances using the 6S atmospheric radiative transfer model and sun photometer data.
PRI from the SE-590 data was formulated:
IPR SE−590=(R569−R529)/(R569+R529) (1)
where R529indicates the reflectance centred at 529 nm
with upper and lower band limits of 536.5 and 521.5 nm (coinciding with the signal of the xantho-phyll cycle) and R569indicates the reflectance centred
at the 569 nm waveband with upper and lower band limits of 561.5 and 576.5 nm (a reference waveband). This differs slightly from that used in previous stud-ies, but still lies within the range of wavelengths considered applicable for the calculation of PRI (Gamon et al., 1993). By referencing R529to R569, this
index normalises for factors which include pigment content and chloroplast movement both of which can affect the R529signal (Gamon et al., 1993).
2.7. A simulation of AVIRIS data
The Airborne Visible Infrared Imaging Spectrom-eter (AVIRIS) (for details see Green et al., 1998) is flown aboard the NASA ER-2 aircraft at an altitude of 20 km and acquires data in 224 contiguous channels of
the shortwave spectrum (400–2500 nm). It measures at a larger spatial scale than the helicopter producing
a ground coverage of∼120 km2 with a 20 m spatial
resolution and spectral resolution of 10 nm. Since this instrument uses new technology that may be amenable for measuring the status of the xanthophyll cycle (but as yet is untested) the data generated by the SE-590 were averaged to those wavebands generated by the AVIRIS instrument (waveband information supplied by Green, NASA, JPL). The SE-590 data were thus averaged to the mid-point of each band, as is standard practice with hyperspectral sensors. PRI was thus cal-culated:
IPR AVIRIS=(R570.5−R530.5)/(R570.5+R530.5) (2)
where R570.5indicates the reflectance over the range of
566–575 nm and R530.5 indicates the reflectance over
the range 526–535 nm. R570.5was the closest available
reference waveband to that used in the computation of PRI from the helicopter data (Eq. (1)).
2.8. Eddy covariance measurement of fluxes
using tower-mounted, eddy covariance systems con-tinuously from May through September of the 1994 growing season. Full details of the theory and in-strument set for each site are published elsewhere. For old black spruce see (Jarvis et al., 1997; old jack pine, Baldocchi et al., 1997; old aspen, Blanken et al., 1997; fen, Sukyer et al., 1997).
2.9. Estimates of stand photosynthesis
The CO2 flux was partitioned into
photosynthe-sis and respiration components by estimating day-time ecosystem respiration as functions of soil tem-perature. Following Goulden et al. (1997), night-time half-hour CO2fluxes at high windspeeds (friction
ve-locity, u*>0.2) were plotted against soil temperature (at 5 cm) and the usual exponential function fitted, of the form,
A=cebTs
where A is carbon dioxide flux , Tsis soil temperature,
c and b are constants and e is the base of the natural
logarithm. Daytime respiration was then calculated us-ing this function and daytime soil temperatures. This was done for each month’s night-time half-hour flux data from May until September for each of the sites (data not shown). Photosynthesis was then calculated as the daytime half-hour CO2fluxes after the
respira-tion component had been removed (Lloyd and Tay-lor, 1994). Because the helicopter overpass occurred on clear bright days (PPFD>900mmol m−2s−1), the
CO2 flux data for these periods were restricted to
PPFD>900mmol m−2s−1. Photosynthesis and inci-dent PPFD were averaged above this PPFD and a daily LUE value computed from these averages ac-cording to the following:
Canopy light-use efficiency
= Canopy photosynthesis
Incident photosynthetic photon flux density
3. Data analysis
3.1. Correlating PRI with LUE data
Relationships between PRI and field measured estimates of photosynthetic LUE were investigated
with linear regression analysis. This was done for the four sites sampled and then with respect to functional group. Relationships between the simulated AVIRIS data and LUE were also investigated with linear re-gression analysis for all species sampled and with respect to functional type.
4. Results
Two mean representative reflectance spectra ob-tained from the SE-590 spectroradiometer, each rep-resenting coniferous or deciduous forest are shown in Fig. 1 along with the two wavebands used for the calculation of PRI.
In the absence of normalisation, the reflectance at 529 nm (R529) yielded no clear correlation with LUE
for the coniferous and deciduous species together (Fig. 2A), or for the deciduous species (Fig. 2B) and conif-erous species (Fig. 2C). However, after normalisa-tion midday top canopy PRISE-590values were
signif-icantly correlated with canopy LUE across the four species sampled (R2=0.64, p<0.05, Fig. 3A). The cor-relations were stronger when the species were divided into their functional groups. A linear relationship was apparent between PRISE-590 and LUE for the
decidu-ous species (R2=0.78, p<0.05, Fig. 3B) and conifer-ous species (R2=0.65, p<0.05, Fig. 3C).
The PRISE-590:LUE relationships were reassessed
by calculating PRISE-590 using a range of reference
wavelengths. Fig. 4A shows a summary of the R2of
the relationship between PRISE-590and LUE. The peak
of this graph would be the best combination of xan-thophyll wavelength and reference wavelength for the estimation of LUE. A reference wavelength of 569 nm produces the best relationship between PRISE-590and
LUE. This is also the optimal reference waveband to use for the deciduous species (Fig. 4B), but for the coniferous sites, the selection of a slightly longer ref-erence wavelength generated a stronger relationship between PRISE-590and LUE (Fig. 4C). In this case,
se-lecting a reference wavelength of 575 nm produced an
R2 of 0.82 in the relationship between PRISE-590and
LUE.
Fig. 1. Mean spectra of the coniferous and deciduous forest canopy from the BOREAS southern study area. Each curve represents the mean of 20–25 scans acquired by the helicopter mounted SE-590 spectroradiometer. Wavelengths used for PRI calculation are shown in the Figure.
relationship across the four species were significantly linearly correlated (R2=0.46, p<0.05, Fig. 5A). The
relationship between PRIAVIRISand LUE was
weak-ened within the deciduous species (R2=0.38, p<0.05, Fig. 5B) but strengthened within the coniferous species (R2=0.78, p<0.05, Fig. 5C).
5. Discussion
5.1. Canopy PRI and photosynthetic LUE
The consistent relationship between PRISE-590and
LUE for canopies sampled in full sunlight (Fig. 3A–C) supports the hypothesis that PRISE-590 provides a
measure of PS II LUE across species and functional types. Similar observations have been made using
individual leaves and small plots (∼2 m) (Gamon
et al., 1990, 1992, 1997; Penuelas et al., 1995, 1997; Fillela et al., 1996), but this is the first time a rela-tionship has been found between PRI and LUE over heterogeneous forest canopies.
Penuelas et al. (1995) studied the relationship between PRI and PS II LUE in plants differing in phenology, habit and photosynthetic type, and found that the slope and intercept of the PRI:LUE relation-ship varied between species. The lack of a single relationship in their study is not surprising. Variation
may arise from the environmental growth conditions, varying anatomy, morphology and pigmentation and diverse species differences. Such factors can dra-matically influence the visible reflectance, and PRI (Guyot, 1990).
The scatter in the relationship between PRISE-590
and LUE could be the result of several factors that can cause divergence between whole leaf assimilation and PS II LUE. These factors include the Mehler reaction, photorespiration (Harbison et al., 1990) and nitrate re-duction (Bloom et al., 1989), all of which compete with carboxylation for reductant generated by electron transport. In the conditions of this study, the signifi-cant correlations between PRISE-590 and LUE within
each functional group suggest that the overall photo-synthetic systems were sufficiently regulated to main-tain consistent relationships between PS II processes and carboxylation.
Fig. 2. Relationship between reflectance at 529 nm and canopy light-use efficiency (LUE). Each point is an average of 20–25 spectral scans and 4 h of canopy LUE data. (A) Four boreal forest species. (B) Two deciduous species. (C) Two coniferous species. Individual species are represented by the different symbols indicated.
Fig. 4. A summary of the R2coefficients for relationship between the photochemical reflectance index (PRI) and canopy light–use efficiency (LUE). (A) Four boreal forest species. (B) Two decid-uous species. (C) Two coniferous species. PRI was calculated us-ing R529as the xanthophyll waveband with reference wavelengths
ranging from 540 to 670 nm, at 3 nm intervals.
some of the scatter in the PRISE-590:LUE relationship
could be attributed to this (Baldocchi et al., 1997).
5.2. Wavelength choice for the calculation of PRI
The wavelength chosen to signal activity of the xan-thophyll cycle differs slightly from that used in previ-ous studies (Gamon et al., 1990, 1992, 1997; Penuelas et al., 1995; Fillela et al., 1996). However, it has been shown that the 531 nm ‘signal’ consists of two spec-tral components (Penuelas et al., 1995; Gamon et al., 1997). Gamon et al. (1997) found the dominance of a 545 nm component in low PPFD, with an appearance of a 526 nm component at higher PPFD. The contri-bution of two spectral components could explain pre-vious reports of a non-linear relationship between PRI and LUE where samples were measured over a wide range of PPFDs (Penuelas et al., 1995). Although Gamon et al. (1990) first highlighted this, a recent pa-per has shown that these two spectral signals are active at different PPFDs, and can be separated in laboratory conditions (Gamon et al., 1997).
The 545 nm component, which was apparent in the difference spectra, is related to chloroplast conforma-tional changes associated with the build-up of the pH gradient, rather than to the conversion of violaxanthin to zeaxanthin. This conclusion was based on exper-imental work associating the generation of the thy-lakoid pH gradient with a change in absorbance at 540 nm (Heber, 1969, Bilger et al., 1989, Bjorkman and Demmig-Adams, 1994) and with a correspond-ing reflectance change near 539 nm (Gamon et al., 1990). That the 545 nm component was initiated at lower PPFDs and progressively disappeared at higher PPFDs, suggests that it is not directly a result of xanthophyll pigment interconversion or the associated heat dissipation, but is more likely a result of light scattering changes associated with the initial build-up of the thylakoid pH gradient upon the sudden transi-tion from darkness to low PPFD (Gamon et al., 1997). The 526 nm component, which appears at higher PPFD, is most likely caused by spectral changes as-sociated with the conversion of violaxanthin to zeax-anthin. This conclusion can be drawn from evidence attributing a similar absorbance change near 515 nm (Bilger et al., 1989; Bjorkman and Demmig-Adams, 1994) and a related reflectance change near 525 nm (Gamon et al., 1990), to the production of
zeaxan-thin. Gamon et al. (1997) further showed that PRI, when calculated using the 526 nm component at high light, was more strongly correlated to LUE than PRI calculated using 531 nm. Thus, the increased
activ-ity at PPFDs above 500mmol m−2s−1 of the 526 nm
component is consistent with the function of xan-thophyll cycle pigments as photoprotective pigments that serve to dissipate excess absorbed PPFD as heat (Demmig-Adams and Adams, 1994; Demmig-Adams et al., 1996). In the linear, quantum yield region of the photosynthetic PPFD response curve, there is no need for such energy dissipation and thus no detectable con-tribution of the 526 nm component in sun acclimated leaves. Given this information, we chose 529 nm to represent our xanthophyll cycle wavelength. However, because no data were acquired at low PPFD, it is not possible to confirm directly the existence of these two spectral components in this study.
We then recalculated PRISE-590 using a range of
reference wavelengths and correlated it against LUE. Some surprising relationships emerged. When all the species were grouped together, the R2 peaked at 0.645 when 569 nm was used as a reference. This was also the case for the deciduous species. However, the PRISE-590 data for the conifers were more strongly
correlated with LUE when a slightly longer refer-ence waveband of 575 nm was used. Gamon et al. (1992) found that any reference waveband between 539 and 570 nm would be suitable for the calcula-tion of PRISE-590and statistical correlation with LUE
data. Their results however, were for a single species, rather than the four species used in this study.
5.3. AVIRIS data and photosynthetic light-use efficiency
suggest that a 10 nm resolution should be appropri-ate for detecting changes in the xanthophyll cycle. Future work using AVIRIS data for PRI applications will confirm its utility for the estimation of canopy scale LUE.
5.4. Effects of background reflectance on PRI
The effective use of periodic remotely-sensed ob-servations in landscape process studies depends on stable estimation of diurnally integrated conditions (Goward and Huemmrich, 1992). It has been widely documented that the reflectance measured from forest sites may vary with solar illumination angle, view-ing angle, understorey or background reflectance and atmospheric effects, as well as intrinsic canopy para-meters such as phytomass and leaf area index of the dominant species (Kimes, 1980; Ranson et al., 1986; Huete, 1987; Guyot, 1990; Goward and Huemm-rich, 1992). Although the spectral data used in this study were acquired from a nadir orientation during conditions of highest possible solar elevation
(ap-proximately 40–60◦), which would have minimised
unstable calculations of PRI, a full sensitivity anal-ysis of the PRI:LUE relationship is warranted and a companion paper will focus on these issues.
6. Conclusions
We have demonstrated that PRISE-590 can serve as
an index of photosynthetic LUE over heterogeneous forest canopies. These results extend beyond earlier studies using individual leaves (Gamon et al., 1992; Penuelas et al., 1995; Fillela et al., 1996) by show-ing that this index is well suited to samplshow-ing areas
of ∼70 m2 in contrasting ecosystems at high PPFD.
Consequently, it should be possible to derive meth-ods of remotely estimating relative photosynthetic rates based on reflectance in the 529 nm spectral region.
Our simulation of AVIRIS data shows that the resolution of this sensor is not too coarse to detect the fine spectral signal generated by the xanthophyll cycle, although it is degraded more at this resolution. Thus we propose that the application at the landscape and larger scales will require a more rigorous test of the use of AVIRIS data as well as the use of radiative
transfer and geometric models, such as SAIL (Scat-tering from Arbitrarily Inclined Leaves) (Verhoef, 1984) and GeoSail (Huemmrich, 1995) to scale PRI and LUE to whole landscapes.
Improvement in the accuracy of scaling processes such as photosynthetic LUE will continue to be piv-otal for the development of physiological ecology and global change research. As new perspectives and methods for scaling ecological function from local to global levels continues to evolve (e.g., via microm-eteorological towers, remote sensing and modeling), our understanding of how functional variables scale across ecological levels must keep pace. Without this synergy, our ability to resolve important issues such as sources and sinks of CO2will be impaired.
Acknowledgements
This work was supported by awards to CJN from the UK’s Natural Environmental Research Council and a National Aeronautics and Space Adminis-tration, Planetary Biology Internship (NASA-PBI), 1998. Eddy correlation fluxes at the old jack pine site were collected by Dennis Baldocchi, Atmospheric Turbulence and Diffusion Division, NOAA, at the fen site by Shashi Verma, University of Nebraska and at the Old Aspen site fluxes were measured by a team consisting of researchers from the Atmospheric Environment Service, Downsview, Ontario and the University of British Columbia, the latter supported by a Collaborative Special Project Grant from the Natural Science and Engineering Research Council of Canada. The 1994 AVIRIS waveband data was provided by Robert Green, NASA Jet Propulsion Lab, California.
References
Asner, G.P., Wessman, C.A., Schimel, D.S., 1998. Heterogeneity of savanna canopy structure and function from imaging spectroscopy and inverse modelling. Ecol. Applications 8 (4), 1022–1036.
Baldocchi, D.D., Vogel, C.A., Hall, B., 1997. Seasonal variation of energy and water vapour exchange rates above and below a boreal jack pine forest canopy. J. Geophys. Res. 102 (D24), 28,939–28,951.
the expoxidation state of xanthophyll cycle components in cotton leaves. Plant Physiol. 91, 542–551.
Bjorkman, O., Demmig-Adams, B., 1994. Regulation of photosynthetic light energy capture, conversion and dissipation in leaves of higher plants. In: Schulze, E.-D., Caldwell, M.M. (Eds.), Ecophysiology of Photosynthesis, Ecological Studies 100. Springer, Berlin, Heidelbeig, New York, pp. 17–47. Blanken, P.D., Black, T.A., Yang, P.C., den Hartog, G., Neumann,
H.H., Nesic, M.D.,. Novak, R., Staebler, R., Lee, X., 1997. Energy balance and canopy conductance of a boreal aspen forest: partitioning overstorey and understorey components. J. Geophys. Res. 102: 28, 915–28, 927.
Bloom, A.J., Calswell, R.M., Finazzo, J., Warner, R.L., Weissbart, J., 1989. Oxygen and carbon dioxide fluxes from barley shoots depends on nitrate assimilation. Plant Physiol. 91, 352–356. Cuenca, R.H., Stangel, D.E., Kelley, S.F., 1997. Soil water balance
in a Boreal forest. J. Geophys. Res. 102 (D24), 29,935–29,365. Demmig-Adams, B., Gilmore, A.M., Adams-III, W.W., 1996. In vivo functions of carotenoids in higher plants. FASEB J. 10, 403–412.
Demmig-Adams, B., Adams, B., 1994. The role of xanthophyll cycle carotenoids in the protection of photosynthesis. Trends in Plant Sci. 1, 21–26.
Fillela, I., Amaro, J.L., Penuelas, J., 1996. Relationship between photosynthetic radiation-use efficiency of barley canopies and the photochemical reflectance index (PRI). Physiol. Plantarum 96, 211–216.
Frank, H.A., Cua, A., Chynwat, V., Young, A., Gosztola, D., Wasielewski, M.R., 1994. Photophysics of the carotenoids associated with the xanthophyll cycle in photosynthesis. Photosynthesis Res. 41, 389–395.
Gamon, J.A., Field, C.B., Bilger, W., Bjorkman, O., Freeden, A., Penuelas, J., 1990. Remote sensing of the xanthophyll cycle and chlorophyll fluorescence in sunflower leaves and canopies. Oecologia 85, 1–7.
Gamon, J.A., Fillela, I., Penuelas, J., 1993. The dynamic 531 nanometer 1reflectance signal: a survey of 20 angiosperm species. Current Topics in Plant Physiol. 8, 172–177. Gamon, J.A., Penuelas, J., Field, C.B., 1992. A narrow waveband
spectral index that tracks diurnal changes in photosynthetic efficiency. Remote Sensing of Environ. 41, 35–44.
Gamon, J.A., Serrano, L., Surfus, J.S., 1997. The photochemical reflectance index: an optical indicator of photosynthetic radiation use efficiency across species, functional types, and nutrient levels. Oecologia 112, 492–501.
Goulden, M.L., Daube, B.C., Fan, S.M., Sutton, A., Bazzaz, J.W., Munger, S.C. and Wofsy S.C., 1997. Physiological responses of a black spruce forest to weather (BOREAS Special Issue). J. Geophys. Res. 102 (D24), 28,987–28,966.
Goward, S.N., Huemmrich, K.F., 1992. Vegetation canopy PAR absorptance and the normalised difference vegetation index: an assessment using the SAIL model. Remote Sensing of Environ. 39, 119–140.
Green, R.O., Sarture, C.M., Chrien, T.G., Aronsson, M., Chippendale, B.J., Faust, J.A., Pavri, B.E., Chovit, C.J., Solis, M., Olah, M.R., Williams, O., 1998. Imaging Spectroscopy and the Airborne Visible/Imaging Spectroradiometer (AVIRIS). Remote Sensing of Environ. 65, 227–248.
Guyot, G., 1990. Optical properties of vegetation canopies. In: Steven, M.D,. Clark, J.A., (Eds.), Applications of Remote Sensing in Agriculture. Butterworths, London, pp. 19–44. Harbison, J., Genty, B., Baker, N.R., 1990. The relationship
between CO2 assimilation and electron transport in leaves.
Photosynthesis Res. 25, 213–224.
Heber, U., 1969. Conformational changes of chloroplasts induced by illumination of leaves in vivo. Biochim. Biophys. Acta. 180, 302–319.
Hogg, E.H., Black, T.A., den Hartog, G., Neumann, H.H., Zimmerman, R, Hurdle, P., Blanken, P.D., Nesic, Z., Yang, P.C., Staebler, R.M., McDonald, K.C., Oren, R., 1997. A comparison of sap flow and eddy fluxes of water vapour from a boreal deciduous forest (BOREAS Special Issue). J. Geophys. Res. 102 (D24), 28,929–28,937.
Huemmrich, K.F., 1995. An analysis of remote sensing of the fraction of absorbed photosynthetically active radiation in forest canopies. Ph.D. Thesis, University of Maryland, USA. Huete, A.R., 1987. Soil and sun angle interactions on partial
canopy spectra. Int. J. Remote Sensing 8 (9), 1307–1317. Jarvis, P.G., Massheder, J.M., Hale, S.E., Moncrieff, J.B., Rayment,
M.B., Scott, S.L., 1997. Seasonal variation of carbon dioxide, water vapour and energy exchanges of a boreal black spruce forest. J. Geophys. Res. 102 (D24), 28,953–28,966.
Kimes, D.S., 1980. Vegetation Reflectance as a function of Solar Zenith Angle. Photogrammic Eng. Remote Sensing 46 (12), 1563–1573.
Lloyd, J.L., Taylor, J.A., 1994. On the temperature dependence of soil respiration. Func. Ecol. 8, 315–323.
Owens, T.G., 1996. Processing of Excitation Energy by Antenna Pigments. In: Neil Baker (Ed.), Photosynthesis and the Environment. Kluwer Academic Publishers, NL, pp. 1–23. Penuelas, J., Filella, I., Gamon, J.A., 1995. Assessment of the
photosynthetic radiation use efficiency with spectral reflectance. New Phytol. 131, 291–296.
Penuelas, J., Llusia, J., Pinol, J., Fillela, I., 1997. Photochemical reflectance index and leaf photosynthetic radiation-use-efficiency assessment in Mediterranean trees. Int. J. Remote Sensing 18 (13), 2863–2868.
Pfundel, E., Bilger, W., 1994. Regulation and possible function of the violaxanthin cycle. Photosynthesis Res. 42, 89–109. Ranson, K.J., Daughtry, C.S.T., Biehl, L.L., 1986. Sun angle, view
angle, and background effects on spectral response of simulated balsam fir canopies. Photogrammic Eng. Remote Sensing 52 (5), 649–658.
Running, S.R., Nemani, R.R., 1988. Relating seasonal patterns of the AVHRR vegetation index to simulated photosynthesis and transpiration of forests in different climates. Remote Sensing of Environ. 24, 347–367.
Sellers, P.J., Hall, F.G., Margolis, H., Kelly, B., Baldocchi, D., den Hartog, G., Cihlar, J., Ryan, M., Goodison, B., Crill, P., Ranson, J., Lettenmaier, D., Wickland, D., 1995. The Boreal-Ecosystem-Atmosphere-Study (BOREAS): an overview and early results from the 1994 field year. Bull. Am. Meteorol. Soc. 76 (9), 1549–1577.
Tenhunen, J.D., Catarino, F.M., Lange, O.L., Oechel, W.C. (Eds.), 1987. Plant Responses to Stress. Functional Analysis in Mediterranean Ecosystems. Springer, Berlin.
Tucker, C.J., 1979. Red and photographic infrared linear combinations for monitoring vegetation. Remote Sensing of Environ. 8, 127–150.
Walthall, C., Williams, D.L., Markham, B., Kalshoven, J., Nelson, R., 1996. Development and present configuration
of the NASA GSFC/WFF helicopter based remote sensing system. Int. Geosci. Remote Sensing Symp., Lincoln, NE.
Verhoef, W., 1984. Light scattering by leaf layers with application to canopy reflectance modelling: the SAIL model. Remote Sensing of Environ. 16, 125–141.
